<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">
<META NAME="Generator" CONTENT="MS Exchange Server version 6.5.7638.1">
<TITLE>Not really what a GIS is for but...</TITLE>
</HEAD>
<BODY>
<!-- Converted from text/plain format -->
<P><FONT SIZE=2>Hello,<BR>
<BR>
I am considering using GRASS to do some astrometric pre-processing of digital astrophotos.<BR>
Of course, this is not what a GIS is for but I'm sure GRASS could help e.g. in the "data reduction" of my images.<BR>
<BR>
For starters, it seems to me that much of the problem of identifying stars in an RGB astrophoto and reducing them to a list of (x,y) points could be tackled both in the spectral and the spatial domains by existing GRASS commands. To start with, I could probably use mapcalc to subtract a dark frame from each channel in the image. The green channel will be brightest but certainly also the noisiest. The red channel will probably suffer from some thermal noise too. And not all stars are blue...<BR>
Well, there must be a way of filtering this all and producing a B/W with clumps which can then be centroided. How would you do that with GRASS?<BR>
<BR>
Any suggestion?<BR>
<BR>
Seb<BR>
<BR>
<BR>
</FONT>
</P>
</BODY>
</HTML>